Pitch Lag Estimation for Speech Concealment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional audio signal processing techniques for speech, particularly in ACELP-like codecs, face challenges in accurately reconstructing pitch information during frame losses, leading to inadequate concealment and pulse resynchronization.
Innovation Solution
The proposed solution involves a weighted pitch prediction method that takes into account the reliability of past pitch lags, using adaptive codebook gains and time weighting to improve pitch lag estimation and pulse resynchronization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If simple repetition based concealment is used, then device complexity is reduced, but pitch reconstruction accuracy deteriorates
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing multiple candidate pitch lags with their corresponding reliability values before concealment is needed. When packet loss occurs, the system can quickly select from pre-evaluated candidates based on reliability metrics, avoiding complex real-time calculations while maintaining high accuracy.
Solution Approach 2:
The patent replaces simple temporal repetition (mechanical copying of last good frame) with a statistical selection mechanism that uses reliability-based weighting. Instead of mechanically repeating the same pitch lag, the system substitutes a probabilistic selection process that chooses from multiple candidates based on their estimated reliability, significantly improving reconstruction accuracy.
2Measurement precision
If pitch extrapolation algorithms are applied, then pitch reconstruction accuracy is improved, but device complexity increases
Solution Approach 1:
The system performs preliminary evaluation of multiple candidate pitch lags and their reliability values during normal operation. This pre-computation allows the concealment algorithm to simply select the best candidate without performing complex extrapolation calculations in real-time, reducing computational complexity while maintaining accuracy.
Solution Approach 2:
Instead of performing complex mathematical extrapolation, the patent copies and evaluates multiple candidate pitch lags that were previously generated by the encoder. The system selects the most reliable candidate based on pre-computed reliability metrics, avoiding the need for complex real-time extrapolation algorithms.
3Reliability
If multiple candidate pitch lags are evaluated with reliability values, then concealment quality is improved, but computational load increases
Solution Approach 1:
The patent shifts computational effort to preliminary action by calculating reliability values and evaluating candidate pitch lags during normal frame processing. This pre-computation allows the concealment phase to operate with minimal computational load, simply selecting from pre-evaluated candidates based on stored reliability metrics.
Solution Approach 2:
The system uses information already available in the encoded bitstream (pitch lag candidates and their associated reliability indicators) to perform concealment. Instead of requiring additional computational resources, the system serves itself by utilizing data that was already processed and stored during normal operation.
Data Source
AI summary
An apparatus for determining an estimated pitch lag is provided. The apparatus includes an input interface for receiving a plurality of original pitch lag values, and a pitch lag estimator for estimating the estimated pitch lag. The pitch lag estimator is configured to estimate the estimated pitch lag depending on a plurality of original pitch lag values and depending on a plurality of information values, wherein for each original pitch lag value of the plurality of original pitch lag values, an information value of the plurality of information values is assigned to the original pitch lag value.